Re: [idn] opting out of SC/TC equivalence

["ben" <ben@cc-www.com>]

Actually, let me clarify what I agree with...

How one feels about this SC/TC equivalence problem really does depend
on which position one takes as decribed by John.

Thanks
Ben Chan

----- Original Message -----
From: "cc-www.com (Ben)" <ben@cc-www.com>
To: "John C Klensin" <klensin@research.att.com>; "Edmon"
<edmon@neteka.com>
Cc: <idn@ops.ietf.org>
Sent: Friday, August 10, 2001 6:51 AM
Subject: Re: [idn] opting out of SC/TC equivalence


Hi John,

I agree with what you said below.

Thanks
Ben Chan

----- Original Message -----
From: "John C Klensin" <klensin@research.att.com>
To: "Edmon" <edmon@neteka.com>
Cc: <idn@ops.ietf.org>
Sent: Thursday, August 09, 2001 11:03 AM
Subject: Re: [idn] opting out of SC/TC equivalence


--On Thursday, 09 August, 2001 10:36 -0400 Edmon
<edmon@neteka.com> wrote:

> Which means that TLD dependent SC=TC should not create any
> problem on a user expectation stand point.
> If you know that typing <SC>.cn and <TC>.cn goes to the same
> place, it will happen no matter where you are connected to the
> network. But that doesnt mean that <SC>.tld and <TC>.tld must
> therefore go to the same place.
> There should be no compelling user expectation for this to
> happen.

Edmon,

Let me say to you --and the WG-- what I said to Dave Crocker
after our exchange of comments at the mike.  It seems to me that
the WG is singing one song, but dancing to a different tune.
There are, I think, two, perfectly reasonable, positions:

(i) The IDN effort is about identifiers and characters,
not about languages or even strings.

(ii) The IDN effort is about making it possible to put
non-English words and phrases into the DNS, just as it is
feasible to put English ones (with some constraints)
there today.

Now, if we really believe the first, then it is not clear how
much of nameprep is needed (somewhat less, I think, although
probably not zero).  But discussions of SC/TC equivalence, Arabic
non-spacing break characters, similarity of characters with
non-similar glyphs between languages, distinctions between CJK
contexts, similar characters in different scripts or blocks, and
so on, are all irrelevant and inappropriate -- those are _all_
conversations about words, languages, and the way language
phrases are written.

But many people --I would suggest a majority of those who care
about the IDN issue, independent of those explicitly represented
in the WG-- are interested in words and languages.  To the extent
that they really represent the requirement for this work, all of
the issues excluded above are important and cannot be excluded.
Your assertion about "no compelling user expectation" is, in that
context, false: a compelling user expectation has been asserted
by CNNIC and others.

And, if any of those language issues are part of the real
requirement, the WG is simply at risk of being irrelevant.  Not
wrong, but irrelevant.   If one reaches that conclusion, then
there is a procedural and responsibility question as to whether
the WG should (and can) report that it can't, within its charter,
do anything relevant or whether it is obligated to report
_something_ out anyway.

    john